UiPath Report: 77% Of IT Executives Plan To Invest in Agentic AI This Year 

Source: Shutterstock

The demand for GenAI has skyrocketed in the last couple of years, but the value of GenAI as a standalone technology may have plateaued, according to a recent survey by UiPath, an enterprise automation and AI software company. Challenges such as integration with enterprise systems, limited autonomy, and governance concerns have slowed its widespread adoption. 

Agentic AI has the potential to bridge the gap by combining AI agents, automation, and people. Unlike traditional GenAI, these intelligent agents can execute complex workflows and interact dynamically with enterprise applications. 

The UiPath report reveals that 92% of U.S. IT executives are extremely or very interested in leveraging agentic AI at their organizations, with 37% reporting that they are already using it. Despite some industry observers arguing that GenAI investments have yet to deliver on their promised ROI, IT leaders remain optimistic. According to the report, 77% of IT executives plan to invest in agentic AI this year 

“I expect that robotic process automation will orchestrate the agents,” said Max Ioffe, Director of the Global Intelligent Automation Center of Excellence at Wesco Distribution. “For larger scale processes, you need clear orchestration and governance, and that means a deterministic technology like RPA (Robotic Process Automation).” 

The IT leaders highlight security concerns, complexity of development, and integration issues as the primary challenges they have encountered when implementing AI in their organizations. With GenAI and LLM solutions specifically, the respondents are most concerned with data quality (47%), IT security risk (33%), and the lack of explainability of results (31%).

How exactly would agentic AI help, and what does this mean for enterprises? 90% of It leaders are confident they have business processes that can be improved by agentic AI. The most appealing benefits of Agentic AI include increased automation (55%), improved problem-solving (53%), and improved accuracy and reduced errors (53%). 

According to Andy Fanning, an industry veteran in automation and AI and CEO of Optura.ai, agentic AI is a natural progression in computing and it represents “another layer of abstraction”, removing the need for constant human input. 

Fanning expects AI agents to “act as experts and orchestrators, determining the best execution methods to complete a task”. He anticipates that agentic AI will thrive in healthcare with the abundance of structured data available for AI to process, however, safeguards would still be needed to build trust and ensure compliance. 

Source: Shutterstock

He further added, “While the agent manages the overall workflow and decision-making, specific actions, like entering or retrieving data from a contract management system, could be handled by RPA bots. These bots would effectively function as tools for the agent, available as needed to support its objectives.”

Agentic AI comes with its own set of challenges. The UiPath report reveals that security (56%), cost (37%), and integration with existing systems (35%) are primary areas of concern. Another challenge is human oversight. Respondents are also concerned about the human element, particularly the need for oversight and upskilling employees to work effectively with agentic AI.

“We use a lot of automation in our processes, and we are moving from providing information to decision support and expert solutions," shared Abhishek Mittal, Global Head for Data Analytics and Operational Excellence, Wolters Kluwer.

“It’s possible that agentic AI could help us do that. But most of our products for customers involve either regulated activities or substantial amounts of money in business transactions. We have generally found that human review is critical, and I don’t see that going away.” 

To overcome the challenges of implementing agentic AI, UiPath recommends starting with small, internal processes that pose minimal financial or security risks. It recommends using orchestration to coordinate agents to ensure governance and keeping a human in the loop for enhanced control and accountability.